Data-Driven Model-Free Predictive Current Control for Brushless Doubly-Fed Induction Generator

The performance of deadbeat predictive current control (DPCC) for brushless doubly-fed induction generator (BDFIG) highly depends on accurate parameters. When the machine parameters utilized in the controllers do not match the actual values, control performance tends to degrade. To overcome this issue, this paper proposes a data-driven model-free predictive current control (MFPCC) method for BDFIG. The proposed MFPCC employs an auto-regressive with exogenous input (ARX) structure to decouple input and output signals while establishing a relationship between them. The coefficients of the proposed ARX structure are designed online using the recursive least square (RLS) algorithm. The results of simulations and experiments demonstrate that the proposed MFPCC not only exhibits good steady-state and dynamic performance when the machine parameters are accurate, but also maintains good control performance when the machine parameters change.